01 / Context
Embed the workflow where it belongs.
A chatbot appears inside a realistic agency site. An assessor works beside the case context. A transaction portal looks like a public service, not a generic SaaS dashboard.
Case study 03 · Public-sector systems
I built an end-to-end state-contract workflow that turns solicitation documents into requirements, architecture, tested demo applications, and a delivery package. Explore the prototypes across corrections, justice, vehicle services, and scheduling.
The problem
A proposal can describe a credible solution and still leave the hardest question unanswered: can this team actually make the workflow work?
State solicitations compress years of domain context into requirements, attachments, security clauses, forms, and evaluation criteria. The challenge is not merely producing screens. It is reconstructing the operational model quickly enough to build something recognizable, defensible, and useful in the pursuit conversation.
The delivery fleet encodes that reconstruction process. It creates a disciplined path from opportunity discovery through demonstration and professionalized package, while keeping human judgment at the teaming, visual, and acceptance gates.
From discovery to handoff
The workflow connects opportunity discovery, RFP analysis, implementation, testing, and delivery packaging. I retain judgment over partner selection, design, and acceptance.
Qualify the solicitation, buyer, deadline, fit, and teaming path before spending build capacity.
Turn the RFP and SOW into workflows, constraints, acceptance criteria, and an evidence-backed solution brief.
Separate architecture, product, data, UI, tests, security, and documentation with explicit ownership.
Deploy the actual workflow, expose meaningful system activity, and test the evaluator journey.
Freeze code, create deployment and security artifacts, map requirements, and prepare the handoff.
Design decisions
A demonstration has to communicate both a credible end-user experience and the engineering underneath it.
01 / Context
A chatbot appears inside a realistic agency site. An assessor works beside the case context. A transaction portal looks like a public service, not a generic SaaS dashboard.
02 / Activity
Retrieval, classification, matching, orchestration, and adaptive behavior become visible through safe evaluator-facing instrumentation rather than unverifiable claims.
03 / Honesty
Evaluation instrumentation is labeled. Capabilities correspond to real behavior. Demonstration proof is never promoted into an award, adoption, or procurement outcome.
Featured demonstrations
Each link opens a working public demonstration retained in the delivery fleet.
Regulated assessment
An assessment workflow designed around structured evidence, review, and controlled decision support.
Public transaction
A citizen-facing transaction and notice experience for a public service context.
Justice workflow
A justice and public-defense experience shaped around institutional users and complex case work.
Institutional scheduling
A scheduling and resource-booking workflow for a public institutional setting.
Voice AI
A supporting demonstration of adaptive, multilingual-aware voice and interview orchestration.
Fleet boundary
These are working demonstrations built against solicitation-shaped problems. They are not presented as deployed agency systems or contract awards.
Outcome
Each delivery leaves behind more than a demo: requirement patterns, visual and technical decisions, packaging standards, tests, and workflow improvements that make the next one faster and more defensible.
The fleet demonstrates breadth across public transactions, corrections, assessment, justice, and scheduling. The deeper signal is the repeatable ability to enter an unfamiliar domain, find its operating constraints, coordinate a build, and prepare the result for another organization to evaluate and own.